For both diagnosis and therapy planning physicians need to access information comprised in medical image data sets, e.g. Computed Tomography scans, stored in a storage system. A tool for retrieval of medical images stored in a storage system is described in an article by H. Mueller et al entitled “The medGIFT project on medical image retrieval”, hereinafter referred to as Ref. 1. The medGIFT tool is strongly based on the GNU Image Finding Tool (GIFT). Like the GIFT, the medGIFT relies on four main groups of features for retrieval: global color features, local color features, global texture features, and local Gabor filter responses. The article also describes different ways of indexing images of lung tissue using clinical information about the contents of images. The article further describes how to remove an image background which does not need to be retrieved. The removal is done by removing specific structures, such as text, followed by a low pass filtering, followed by thresholding and removal of small unconnected objects.